21

Concrete Repair, Rehabilitation and Retrofitting II – Alexander et al (eds)© 2009 Taylor & Francis Group, London, ISBN 978-0-415-46850-3

Advanced NDT methods for the assessment of concrete structures

H. WiggenhauserFederal Institute for Material Research and Testing (BAM), Berlin, Germany

ABSTRACT: The development of methods for the Non-Destructive Evaluation (NDE) of concrete and masonry structures has made considerable progress in the past two decades. Technology and knowledge transfer from other areas such as materials testing and medicine has been the main source for innovation, supported through multidisciplinary research by scientists, technicians and support personnel. A versatile toolbox of methods for the investigation of RC structures has emerged from this effort. In the area of simulation and 3D-reconstruction of ultrasonic and radar data, progress was driven by more powerful computers and development of state of the art software for 3D re-construction and visualization of data.

Bringing these developments into the state of practice, into everyday use on construction sites by trained engineers rather than researchers still remains a major challenge. To earn the trust of industry, the reliability of these Non-Destructive Testing (NDT) methods should be established and adequately verified to prove that their application is worth the additional expense and effort.

Validation of NDT methods has to prove that the customer requirements are met by the test. A focused effort is needed to implement a validation methodology for NDT methods in civil engineering, which is widely accepted and used. In an effort towards this goal, a number of reference specimens have been developed over the years at BAM and other locations, acting as a standard for specific testing tasks. Those specimens are available to the NDT community for comparing their methods and instruments under independent conditions.

The development and application of NDT methods for concrete structures requires more than just using off-the-shelf testing equipment. Existing test methods must be improved for more reliable use. New methods borrowed from other areas of testing such as medicine or geophysics should be further developed. Using an array of methods and combining their results, preferably through data-fusion algorithms has a lot of potential. Over the years, the NDT group at BAM has developed and accumulated valuable expertise in some of the above-mentioned areas.

The focus of the research and developments efforts at the NDT-CE division at BAM has been on methods used to evaluate the geometrical properties of structure such as thickness, location of components as well as detection of the presence of voids, delaminations, ungrouted tendon ducts or cracks.

the concrete, 12 transducers each are combined and form a large aperture for transmitting and receiving the signals.

Scanning of lines and areas was possible with the A1220 sensors in reasonable time, due to the fact that no special coupling was required. With the scan-ning systems developed at BAM, automated testing utilizing these transducers became a reality. Based on these experimental developments, reconstruction processes developed in other areas could be utilized for ultrasonic data collected within a 2D aperture. The result of such tests is a 3D reconstruction of the volume below the test surface (Streicher et al. 2005). As an example of 3D imaging of concrete bridges a C-scan from a bridge is shown in Figure 2. It clearly shows tendon ducts in a depth of 10 cm over the entire length (10 m) of the tested web.

1 ULTRASONIC PULSE ECHO (UPE)

It became clear very early that ultrasonic pulse echo on concrete requires very low frequency transducers (Neisecke 1991) to overcome the scat-tering problem in coarse aggregate concrete. Multi position experiments and synthetic aperture data analysis led to reliable thickness measurements of concrete slabs. In order to reduce testing times and to ease experiments, arrays of ten transducers were used. In the mid 90s a new type of transducer was developed in Russia (www.acsys.ru 2006) which did not require special coupling to the concrete surface (Fig. 1). Based on transversal excited ultra-sound, testing became much easier to perform with the availability of these commercial products. The individual transducers have ceramic pin contact to

22

The 3D reconstruction of ultrasonic echo data is based on the well known SAFT (Synthetic Aperture Focussing Technique) algorithm which calculates the phase shift for each single measurement for each volume element. Calculation times have dropped dramatically due to faster processors and advanced programming techniques. Quasi real-time imaging is becoming a reality.

Very recently a new ultrasonic device has become available which allows taking measurements and imaging a scan within seconds. Based upon the same point contact transducers, measurement time could be reduced to increase performance. With this unique device, quasi real-time imaging of concrete using a hard-held device is available now (Fig. 3).

The coupling problem for ultrasound can be mini-mized using air coupled transducers. In transmission experiments this technique has shown its capabilities (Hillger 2004). However, in pulse echo mode, experi-mental obstacles are apparent. The separation of the echo signal from all other signals (surface echo, direct coupling between transmitter and receiver, surface wave) is quite difficult. Further research is needed to utilize this promising technique. The advantage

Figure 1. Ultrasonic device with point contact 55 kHz shear wave transducers. 12 point contacts used for transmis-sion of ultrasound, 12 for receiving (ACSYS).

Figure 3. Ultrasonic array for real time imaging. Ten Elements are aligned to form a linear aperture. B-scan imaging within 4 seconds. Shear wave transducers, 55 kHz. (www.acsys.ru)

Figure 4. Air coupled ultrasonic probes mounted in front of concrete block for pulse echo measurements. Aluminium sheets mounted for air wave suppression (BAM).

Figure 2. Result from ultrasonic investigation of PT con-crete bridge. Shown is the web with tendon ducts located in depth range around 10 cm. Darker colors indicate higher acoustical contrast.

23

of this technique is the superior testing speed which could be used for testing large areas in relatively short time. Figure 4 shows the experimental setup for testing concrete with air coupled ultrasound in pulse echo mode.

2 IMPACT-ECHO (IE)

Impact-Echo is an acoustical test method based on mechanical excitation of sound waves in concrete (Sansalone & Streett 1997, Pessikki & Olson 1997). Resonant vibrations of the surface near the point of contact are recorded and identified by transforming the measured time signal into the frequency domain (Fig. 5). Dominant frequencies are assigned to depth values by applying the well known IE formula

d v fL= / 2 (1)

with vL being the P-wave velocity and f the meas-

ured frequency. The wave speed must be determined at each concrete through calibration at a position of know thickness or by measuring on a core.

Impact-Echo has been successfully applied to con-crete thickness measurements and defect localization in many cases. For this purpose a number of com-mercially available Impact—Echo devices are avail-able (Figs. 6, 7). These devices are for manual point testing and can be distinguished by their impactors

Figure 5. Impact-Echo signal: Time domain data (top) and frequency power spectrum (bottom). Thickness signal at 6.81 KHz, scanner resonance at 1.2 KHz.

Figure 6. Impact-Echo thickness gauge by Olson Instru-ments (right). Impact is generated by a solenoid, coupling to the surface through elastic material. Frequency range: 1–25 KHz. Rolling JE probe shown on the left. Impact is generated every inch (2.54 cm) for continuous measure-ments. (www.olsoninstruments.com)

Figure 7. Impact-Echo system of Germann. Impact gener-ated manually with steel balls. Coupling to surface through thin lead sheets. Frequency range: 1–50 KHz. (www.germann.org)

and sensors. Sensor coupling is assured through lead caps or plastic material, impacts are generated either through hand held steel balls of different diameters or by an electrically driven solenoid. Latter has the advantage of being easily reproducible when the IE head is not moved between subsequent impacts.

24

Figure 8. Thickness of concrete slab of 5 × 4 m2 (BAM reference specimen GBP. Top: plan of section with thickness variations. Middle: 3D plot of measurements in very dense grid (2 × 2 cm²). Bottom: B, C and D Scan of same measure-ments with signals of various thicknesses.

At BAM this point method was improved into a scanning test method to visualize test results as an Impact-Echogram, similar to a B-scan in ultrasonic pulse echo or a GPR radargram. The amplitude of the frequency power spectrum is displayed colour coded or in greyscale over the position of the measurements and frequency (Fig. 8).

A handheld IE test device for scanning measure-ments was developed by Olson Instruments (www.olsoninstruments.com) as shown in Figure 6. A self driven scanner for horizontal surfaces was developed at BAM for both, Impact-Echo and Ultrasonic sensors (Fig. 9).

With automated testing in dense measurement grids Impact-Echo measurements can show very detailed the thickness of a concrete slab. When scan-ning is extended to a 2-dimensional grid, IE-results can be shown as 3-D images as shown in Figure 8.

It was successfully shown, that grouting defects in tendon ducts can be identified by the apparent increase in thickness of the concrete slab at the posi-tion of the duct. However, validation of this test is still pending to define the limits of this test method and making it a reliable tool in quality control (Fig. 10).

In a recent study at University Karlsruhe the reli-ability of Impact-Echo for the localization of honey-combs in concrete slabs has been researched (Müller et al. in prep.). It could be shown, that Impact-Echo can directly and indirectly determine the presence of a honeycomb. The indirect method identifies a defect by the missing thickness signal. However, it should be clearly pointed out, that in this study the honeycombs could only be reliable located with scanning Impact-Echo. Point measurements were not showing reliable results (Fig. 11).

Impact-Echo experiments on specimens with small dimensions may be interpreted with care. Since most of the impact energy is transformed into surface and shear waves, reflections of those from nearby edges

Figure 9. BAM NDT Stepper FOR Impact-Echo and Ultrasonic Pulse Echo.

The diameter of the steel balls determines the fre-quency content of the impact, given by the formula

f dmax /= 291 (2)

with d being the diameter of the steel ball in mm and fmax

being the maximum useable frequency in kHz (www.olsoninstruments.com 2006). The duration of the impact ultimately determines the maximum fre-quency of the impact. This equation may only be used for concrete with perfect surface. In practical cases, the surface must be checked for dirt and soft areas. The impact will generate different waves depending upon the exact position where the hit takes place. If the impactor directly hits a hard aggregate, which may be hidden by a thin cement paste cover, the frequency range may be different from a hit into the softer cement paste.

25

attenuated. With an AD-converter of low dynamic range (e.g. 8 bit) the signal amplitudes may not be recorded properly and make the data less accurate.

Data analysis of Impact-Echo signals includes time and frequency domain filtering. Typically a high pass filter is applied to remove the signal offset and low frequency resonances within the system. A time band pass excludes signals from times where the Impact-Echo signals should have attenuated. If the time band pass is set too small, frequency resolution is lost. Wavelet analysis has been used for data analysis as well and can serve very well for time-frequency anal-ysis. Recently, the Hilbert-Huang transform has also been utilized for the successful analysis of Impact-Echo signals (Algernon 2006).

3 GROUND PENETRATING RADAR (GPR)

Ground Penetrating Radar has emerged from geo-physical applications and has become a versatile and powerful instrument for civil engineering applica-tions. GPR can be applied in general without surface preparation by sliding the antenna over the surface. The fast data collection allows covering a large area in reasonable time.

Due to the nature of electromagnetic waves it can be used for concrete and masonry structures but radar waves cannot penetrate metals. Any metallic layer, e.g. a metal sheet or a metallic tendon duct is an inpenetratable boundary (Fig. 12). However, rein-forcement which leaves gaps between the rods can be passed by radar waves within limits.

The propagation of electromagnetic waves in solids depends on the dielectric constant (relative permittiv-ity) of the material. The permittivity of the material basically describes how the dielectric field in the mate-rial is following the applied electric field of the waves. The permittivity is a complex number. The real part determines the propagation velocity in the material

v cL r= 0/ ε (3)

Figure 10. Impact-Echo testing of grouting defects in ducts. Top: drawing of defect location. Middle: B-scan along duct. Bottom: C-scan showing apparent increase of slab thickness at defect locations. (BAM, unpubl.)

Figure 11. Honeycombs in concrete slabs measured with scanning Impact-Echo. Left: Pre-fabricated honeycombs of different sizes. Top: Drawing of specimens with differ-ent reinforcing ratios. Middle: Pictures of specimens before casting. Bottom: Impact-Echo results, darker colour indicat-ing larger thickness. Thickness is apparently increased over honeycombs.

cause interference patterns on the surface. Such signals can easily being mistaken for thickness sig-nals or void indication if not validated through meas-urements at different points: Those must be separated well enough to exclude that the signal is triggered by geometry effects.

Before testing, the operator has to assure that the system setup is meeting the requirements of the test. The sampling rate of the data acquisition system must be high enough to measure the highest envisaged fre-quency (Nyquist frequency). Also, the impactor must be able to generate the frequency for the given test problem. The sensor must work accordingly and the data acquisition system must record the signal prop-erly. The dynamic range of the AD-converter becomes an issue when the initial signal after the impact is very high but the following resonances are highly

Figure 12. Radar parabola signature when passing over reflector.

26

with c0 being the velocity of light in vacuum. The

imaginary part is largely responsible for the frequency dependent attenuation of the radar waves.

Radar waves are polarized, having the effect of a preferred direction. This is especially true for linear objects like reinforcing bars, which show up with highest intensity when the polarization of the anten-nae is parallel to the direction of the rebar. For this reason the localisation of rebars should always be per-formed using both polarization directions (Fig. 13). This is done by scanning a surface in two perpendicu-lar directions and combining the two data sets.

The availability of antennas with different fre-quencies allows choosing the penetration depth in accordance with the requirements of the test. Low frequency antennas penetrate deeper into the mate-rial, but have lower depth resolution. High frequency antennas have a higher depth resolution but less pen-etration depth.

For application on structures mostly bow-tie anten-nae types are used. They can be slid over the surface for conducting the test. Horn antennas are mainly being used for road and train track measurements. They have a distance of several decimeters from the surface and can be mounted on vehicles for fast inspection.

Typical applications for radar on concrete struc-tures are (Maierhofer et al. 1995):

− Location and concrete cover of reinforcement− Location and concrete cover of tendon ducts− Thickness measurements of beams and slabs− Moisture detection areas− Localization of voids− Localization of delaminations− Measurement of the thickness of layers

Conducting radar tests and the interpretation of the results is supported by advanced instrumentation which is commercially available. More important are

powerful computer systems which can handle large amounts of data in reasonable time. Software tools for different purposes have been developed and are available for service providers or research.

GPR applications for non-destructive testing in civil engineering have been a success story for many years now. More and more information has become available and known about the capabilities and limita-tions of such systems.

New systems emerging from research will bring array systems into the market, allowing to scan a wide path with one sweep. An early system of this type has been developed by the Federal Highway Administra-tion in the nineties (Scott et al. 2000).

4 THERMOGRAPHY

Impulse Thermography is widely used in non-destructive testing of composite materials, such as aircraft components. In civil engineering, thermog-raphy is mostly used in the quasi static mode, where the ambient temperature distribution is utilized to test an object. Impulse heating of a surface, followed by a recording and analysis of the temperature develop-ment over time and position leads to new applica-tions. In recent studies, impulse thermography has been applied to carbon fiber reinforcements and shal-low defects in concrete with good success. Figure 14 shows a thermogram of a concrete specimen with carbon fiber reinforcements which has some artifi-cial defects. The temperature decay curves show a distinct difference between sound and voided areas (Helmerich et al. 2006).

The quantitative analysis of the temperature after heating can either be done in the time domain by curve fitting or in the frequency domain. Later anal-ysis is referred to as Pulse-Phase-Thermography and has certain advantages. Surface effects are largely eliminated in the PPT images, from the phase angle the depth of defects can be derived(Arndt 2006).

Figure 13. Influence of polarization on the detectability of tendons.

Figure 14. Thermogram and temperature decay curves of carbon fiber reinforcements after impulse heating (1 min, 7.2 kW/2.25 m2.

22.0

29.4

Tem

pera

ture

in o C

0 20 40 60 80 100 120 140 16022

24

26

28

30

32

34 infrared radiator

sound area disbonded area difference between disbonded and sound area

Time in s

Tem

pera

ture

in °

C

0

1

2

3

4

Tem

perature difference in K

27

5 TOMOGRAPHY

Tomographic investigation of concrete and masonry structures is time consuming and complicated. X-ray tomography is not very common for large objects, acoustic tomography has been tried and results have been reported (Colla et al. 1997, Patitz & Illich 2000). New sensors (ACSYS ultrasonic device A1220 with dry coupling) and the availability of automated measurements with the scanner systems allow new approaches to tomography of large civil engineering structures. A more general approach of tomography is envisaged for this research, where large sensor arrays record signals which are transmitted or reflected. The combination of methods (radar, ultrasonic and sonic) allows a reconstruction of the volume under investi-gation on a much larger data basis than from manual testing with a single method. Figure 15 shows the results of tomographic measurements and inversion of the data on a masonry specimen.

6 LIBS

The application of the Laser-induced Breakdown Spectroscopy (LIBS) for the quantitative evaluation of chlorine and sulphur content in building materials is being researched at BAM (Fig. 16).

Figure 15. Result of acoustic tomography of a masonry specimen with two voids (missing bricks). Forward mod-eling: acoustic waves. Transmitter T and Receiver R every 5 cm. Bricks/cement: v = 3000 m/s, ρ = 1.7 g/cm³, Voids: v = 300 m/s, ρ = “1” g/cm³, Inversion: 5 × 5 cm grid size, modified velocity model, five iterations, curved rays.

Pos

ition

of T

rans

mitt

er [c

m]

T1R1 0

20

40

60

80

100

0 20 40

0.25

0.26

0.27

0.28

0.29

0.3

0.31[10^2 m/s]

R23 T23

Pos

ition

of R

ecei

ver

[cm

]

Figure 16. Experimental LIBS set-up in the BAM laboratory.

Figure 17. Top: Typical spectrum of cement measured under helium containing chlorine, the spectral lines are labelled. Bottom: Comparison of spectra of cement and aggregate.

LIBS is a method which has the ability to evaluate the elements of the concrete matrix to get informa-tion of the bulk material and at the same measurement also information about trace elements. The method can be applied on surfaces without any preparation and the results are available directly after the meas-urement. LIBS is currently developed as a tool for on-line evaluation and for quality assurance during repair works. LIBS uses a pulsed infrared laser for

28

plasma generation and a spectroscopic system for the analysis of the emitted plasma radiation.

For the quantitative evaluation of chloride ingress the chlorine spectral line at 837.6 nm is used. The sul-phur spectral line is for sulphate ingress at 921.6 nm. For every measurement the ratio of a calcium to an oxygen spectral line is evaluated to take care for the heterogeneity of the concrete and relate the results directly to the cement content (Fig. 17).

7 ULTRASONIC ECHO AND IMAGING TECHNIQUES

Foundations have only one-sided access and require echo methods like ultrasonic echo (US-echo) for test-ing. The investigation of the above-mentioned slab has been carried out with an array of 24 dry point contact transducers working with transverse waves excited with a centre frequency of 15 kHz (ACSYS A1220).

For a reliable imaging of the complex geometry an automated transducer positioning system (scan-ner) has been used. Data (8000 measuring points) have been recorded and processed with the help of a reconstruction calculation. The so-called SAFT (Synthetic Aperture Focusing Technique) focuses signals received at many aperture points by coher-ent superposition, yielding a high-resolution image of the region of interest. The following images are results of SAFT-reconstruction. Various sec-tions through the reconstructed data volume can be processed and layers with significant reflections

become obvious and visualise internal objects and geometry.

Figure 18 shows the section parallel to the surface at a depth of 75 cm. The expected reflection of the back wall at that depth is clearly visible. In upper left and upper right corner (x/y 960/860 mm, x/y 4000/850 mm) two small areas show no back wall reflection. This is where the pile heads are located because the signals propagate from the slab further into the piles and are not reflected at the depth of the back wall.

The limits of the thickness measurement depend-ing on the reinforcement ratio become clearer look-ing at cross-sections shown in Figure 19. For the high reinforcement ratio shown at the top of Figure 20 (section b-b) only in the non-reinforced sections a clear back wall signal at depth of 75 cm and 125 cm appears. In the sections with upper and lower rein-forcement a strong reflection in the surface near depth is visible. A weak signal at the 75 cm section allows depth estimation up to that depth. If the 28 mm diam-eter reinforcement is placed only in the lower level, as shown in section c-c at the bottom of Figure 19, the reinforcement bars produce a reflection in addition

Figure 18. Section parallel to the surface at depth of 75 cm.

Figure 19. Drawing of the designed foundation slab with varying thickness and reinforcement ratio.

Figure 20. Section b-b, ∅ = 28, s = 10 cm (upper and lower reinforcement).

29

to the back wall reflection a few centimetres above. Depth measurement up to 125 cm is possible with this reinforcement ratio.

Cross-section d-d shown in Figure 21 also reveals the geometry of the slab and the location of the piles. The back wall reflection at 75 cm and the bot-tom of the strip foundation at 125 cm are clearly visible. Also the detected width of the strip founda-tion of 50 cm and its location agree well with real-ity. The interrupted back wall echoes at the depth of 75 cm between x = 800 mm and 1100 mm and x = 3850 mm and 4150 mm mark the location of the pile heads.

Reliable thickness measurements have been pos-sible up to 75 cm for areas with high reinforcement ratio (28 mm diameter at 100 mm spacing, cross-wise, upper and lower reinforcement layer) and up to 125 cm with lower reinforcement ratio.

8 AUTOMATION

Automation of non-destructive testing methods helps to overcome the disadvantages of manual point test-ing. Point measurements must be repeated several times for statistical proof of the results. Reproduc-ibility is not always given in rough test environments and difficult surfaces. Furthermore test personnel get tired during long testing periods.

At BAM several automation devices have been developed for various applications: automation of point testing methods with physical contact dur-ing testing (impact-echo, ultrasonic echo) and those with non-contact testing (radar, rebar detection). The test devices are mounted on a two-axial frame which moves to a pre-defined position and triggers the measurement. Pneumatic installations are used to couple the sensors to the surface until the test is fin-ished. The entire test procedure is computer control-led. All data about the test procedure, the test object and the applied methods is stored for documentation. Interface with test equipment is done by utilizing existing interfaces or in cooperation with manufactur-ers. Typical test times required for ultrasonic testing is 1.5 h per square meter at a spacing of 2 cm × 2 cm

Figure 22. BAM NDT-Scanners for concrete surfaces. Top: Vacuum scanner 1.2 × 1.2 m². Middle: Horizontal sur-faces 10 × 4 m². Bottom: Vertical surface 10 × 16 m².

Figure 21. Cross section d-d revealing the slab geometry and location of the pile heads.

30

between test positions. Radar is much faster due to the non-contact nature of the test, at a spacing of 5 cm between lines the required time is 5 min per square meter.

Scanners have been developed for both horizontal (bottom and ceiling) and vertical flat surfaces. The large scanner covers an area of 10 m × 4 m, smaller versions exist for 1.8 m × 1.8 m areas. Recent suc-cessful tests have supported the concept of vacuum mounting of a lightweight frame. This version does not require fixed mounting to the concrete surface (Fig. 22).

Mounting and test times have been successfully reduced using lightweight frames and fast, network based programs which can be easily adapted to test-ing scenarios. Fully automated testing of a 10 m² area concrete area within a day is within reach. So far the biggest problem is accessibility, which should be eased by the vacuum attached lightweight scan-ner which can be mounted from a cherry picker of snooper truck.

For horizontal testing the “BAM NDT-Stepper” was constructed (US Patent 2006/0191358 A1). A small vehicle moves with constant speed along a straight line. The sensor is pneumatically pressed to the surface in equally spaced positions along this line. The test position is maintained exactly during the test while the vehicle is moving. Spacing can be as long as 10 cm, the speed is defined by the spacing and the test time required per position (typical speed is 1–3 min/m). Air pressure is generated by an internal miniature pump. An internal battery allows testing time of more than four hours for typical test scenarios. (Fig. 23).

New versions of the “BAM NDT-Stepper” include mounting of two sensors on either side of the vehi-cle. Array systems can be mounted on an extension

of the mechanics which assures the positioning of the sensors.

9 BRIDGE TESTING

Non-destructive assessment of existing bridges is the most important application of the NDT developments at BAM. Bridges are a valuable infrastructure asset. Germany alone has app. 37,000 rc-bridges within in the highway system. Interruptions to the flow of traffic should be minimized for inspections and maintenance. In addition to the visual inspection based regular inspec-tion, special testing is required in some cases where damages are possible but not identifiable visually.

For prestressed concrete structures the investi-gation of tendon ducts is one of the most essential test problems. The localisation of tendon ducts, the determination of the thickness of the concrete cover and especially the detection and quantification of ungrouted areas inside the ducts are relevant ques-tions, which can be solved by the combined applica-tion of the imaging echo methods radar, impact-echo

Figure 23. BAM NDT-Stepper with mounted ultrasonic sensor (ACSYS A1220).

Figure 24. Visualization of the outer reinforcement layer of a box girder web from radar data. Measurements were taken in two perpendicular polarization directions, data was fused and reconstructed.

Figure 25. Arrangement of tendon ducts in the cross sec-tion of a box girder web. Left: According to construction plan. Right: Located at a SAFT-B-projection by ultrasonic echo.

Mea

sure

men

t sur

face

31

and ultrasonic echo. Other points of investigations are detecting the reinforcement and determining the thickness of structures if accessibility is given from one side only.

Investigations with the different test methods utilizing the scanning systems were carried out on several post-tensioned concrete bridges. The improvement of the measurement conditions and the more comprehensive data acquisition allowed a much simpler interpretation of the results. Rein-forcement bars, which are arranged perpendicular could be visualised with a high resolution (Fig. 24). Tendons in the investigated bridge structures could be localised with the 1.5 GHz-antenna at measure-ment depths up to 16 cm and with ultrasonic echo at measurement depths up to 55 cm (Fig. 25). The position of couplings between tendons could be determined and hints to not completely filled ten-don ducts could be given. The backside of structures were detected with ultrasonic and impact-echo up to a depth of 83 cm.

The scanning systems show a reliable performance under field conditions. The time and effort could be considerably reduced by automation in comparison with manual measurements (Streicher et al. 2005).

10 VALIDATION

Validation is used in many technical fields, e.g. phar-macy, aviation, to furnish the technical proof that the customer’s requirements are fulfilled. Subject of the validation is the whole process of a successful solu-tion for a testing problem. Well-documented valida-tion improves the application safety for a service provider and makes his service transparent for the client. The client knows exactly what he buys and that he has full benefit from the results.

According to DIN EN ISO 17025 (DIN EN ISO/IEC 17025, 2004) Validation is the confirma-tion by testing to furnish the proof that the require-ments for a certain intended use can be fulfilled. The client—in some cases together with the service pro-vider—expresses his/her requirements considering all necessary boundary conditions. If a testing method—suggested by the service provider—succeeds to satisfy these requirements the validation process is completed. According to Figure 26 the validation process consists of three steps:

− Characterization of a testing methods− Requirements of the client− Proof that the client’s requirements have been

fulfilled

Figure 26 also shows this process: characterization is done by regular testing in research and development

and evaluating the results. Calculating the uncertainty of measurement, determining the precision and accu-racy and limit of detection etc. are possibilities to char-acterize a method under certain boundary conditions (method, device, environment etc.). The customer requirements are also expressed by the properties mentioned before. They are influenced by time, costs, accessibility etc. All this together forms the indi-vidual testing problem. A testing method under the characterized methods that is suitable for the testing problem is marked by the intersection of characteri-zation and customer requirement. The validation is at the end of the process if the proof is furnished that customer’s requirements are satisfied.

This does not mean that a validation always has to be carried out at the customers testing object because this will often be impossible. The idea of the process is to provide a close characterization of the method considering a wide range of boundary con-ditions that fit for the customer’s testing problem. In one case it might be possible to predict precisely that the customer needs will be satisfied. In another case the results, e.g. the expected uncertainty, might be too close to the customer’s defined limit and the success of a test cannot be guaranteed. In case that

Figure 26. Validation process.

Figure 27. Total standard deviation of thickness measure-ment with ultrasonic-echo (transverse waves) in relation to thickness, reinforcement and data assessment.

32

the characterization of a method is not sufficient, further—but few—measurements considering the unknown boundary conditions have to be carried out. Validation has to be carried out individually for a spe-cific testing problem because there is obviously no general validation for a testing method.

The Guide to the Expression of Uncertainty in Measurement (GUM) (Guide to the Expression of Uncertainty in Measurement 1995) provides a good methodology that can be applied in the NDT-CE field. According to the GUM the uncertainty of thickness measurement of 70 to 120 cm thickness and having areas of different reinforcement ratio was measured. The testing was performed with an automated test-ing device using dry point contact sensors (ACSYS A1220).

The measured data was first assessed as raw data and later assessed after improving it by creating the envelope of the transit time curve. In Figure 27 the effect of improved data analysis for different rein-forcements is shown. We assess the data according to the GUM mentioned in the previous chapter the total standard deviation for the thickness measurement in relation to thickness, reinforcement and data assess-ment can be calculated (Gaussian error propagation) as given in Figure 27.

The results shown in Figure 27 are the results of a systematic characterization. It helps the engineer to estimate up to which depth and to which reinforce-ment he may expect reliable results. If the engineer just needs a quick overview to decide if thickness is 70 cm or 120 cm he/she may be satisfied by assessing raw data. If he/she reliably wants to know the slab thickness, e.g. in case of a static calculation, he/she will get results with a standard deviation less than 5% only by using “improved data” (in this case creating the envelope) and up to a reinforcement of ∅ 12 mm with 15 cm spacing (upper and lower layer).

11 NDT-CE COMPENDIUM

Information about the availability and performance of NDT-CE methods is not readily available. The BAM in Germany has compiled a compendium on such meth-ods (ZFPBau-Kompendium) and published it later in the internet: www.bam.de/zfpbau-kompendium.htm. The English translation is under way and will then also be freely available worldwide on the same address.

12 REFERENCE SPECIMEN

Information Improvements of NDT methods can best be shown through reference tests at suitable specimens. For this purpose, a number of large scale

reference test specimens have been constructed at BAM. Well defined test problems are incorporated into those, well separated from other defects. Special care is taken of size effects, e.g. the geometry effects in impact echo. The specimens typically have a size of at least 2 m × 1.5 m × 0.5 m. The size is mainly limited by the ability to move the specimens with a fork lift.

In addition to the laboratory specimens, even larger specimens have been established for special tasks:

− The large concrete slab (LCS) with test problems typical for bridge decks (Figs. 28, 29)

− Foundation slab with different reinforcements, thickness sections and test piles underneath

All test specimens are available for round robin tests to other interested groups upon request.

Figure 28. Large concrete slab on BAM campus (10 m × 4 m, thickness 30 cm).

Figure 29. Large concrete slab during construction: rear section with tendon ducts, front section with thickness vari-ations and honeycombs.

33

ACKNOWLEDGEMENTS

The research described in this contribution is the pri-mary result of ongoing teamwork in division VIII.2 of BAM. The dedication and enthusiasm of all members is greatly appreciated. Special thanks to all research partners within national and international projects.

REFERENCES

Algernon, D. 2006. Analyse akustischer Wellen in Beton. Berlin: Dissertation Technische Universität Berlin.

Arndt, R. 2006. Rechteckimpuls-Thermografie im Frequenz-bereich. Berlin: Dissertation eingereicht bei der TU.

Colla, C., Das, P., McCann, D. & Forde, M. 1997: Sonic, electromagnetic and impulse radar investigation of stone masonry bridges. in: NDT&E International Sonderheft Vol. 30 4: 249–254.

DIN EN ISO/IEC 17025: Allgemeine Anforderungen an die Kompetenz von Prüf- und Kalibrierlaboratorien (ISO/IEC 17025:1999); Ausgabe: 2000-04, Beuth Verlag, Berlin.

Guide to the Expression of Uncertainty in Measurement: deutsche Übersetzung 1995. Leitfaden zur Angabe der Unsicherheit beim Messen, Beuth-Verlag, Berlin.

Hasenstab, A., Krause, M., Rieck, C. & Hillemeier, B. 2004. Materialuntersuchung an Holz mit niederfrequenter Ultraschall-Echotechnik, Berichtsband 89-CD, DACH—Jahrestagung 2004, 17–19. Mai 2004, Salzburg.

Helmerich, R., Niederleithinger, E. & Wiggenhauser, H. 2006. Toolbox with Nondestructive Testing Methods for Condition Assessment of Railway Bridges; Trans-portation Research Record: Journal of the Transporta-tion Research Board, no. 1943, Transportation Research Board: 65–73.

Hillger, W. 2004. Imaging of Defects in Concrete Compo-nents with Non-Contact Ultrasonic Testing, 16th World Conference on NDT, 30. August–03. September 2004, Montréal, Canada.

Maierhofer, C., Borchardt, K. & Henschen J. 1995. Appli-cation and Optimization of Impulse Radar for Nonde-structive Testing in Civil Engineering and further articles therein, Proceedings of the International Symposium on Non-Destructive Testing in Civil Engineering, Deutsche Gesellschaft für Zerstörungsfreie Prüfung e.V., Eds: Schickert, G. & Wiggenhauser, H., Berlin: 663–672.

Müller, H.S., Fenchel, M., Wiggenhauser, H., Maierhofer, C. & Krause, M. in prep.: Zerstörungsfreie Ortung von Gefügestörungen im Beton, DAfStB Forschungsbericht.

Neisecke, J. 1991. Ultraschallprüfung im Bauwesen—eine kritische Wertung, in: Schickert, G. (Hrsg.); “Zerstö-rungsfreie Prüfung im Bauwesen”, Tagungsbericht Int. ZfPBau-Symposium 27. February—1. Macrch 1991 in Berlin, BAM, DGZfP, Berlin: 182–191.

Patitz, G. & Illich, B. 2000. Blick ins Innere: Zerstörungs-arme Untersuchungen: Voraussetzung für ein substanz-schonendes Sanierungskonzept/Teil 2, Bautenschutz + Bausanierung, Vol. 23, no. 4: 35–39.

Pessikki, S. & Olson, L. 1997. Innovations in Non-Destructive Testing of Conncrete, Americam Concrete Instituite, Farmington Hills, SP-168.

Sansalone, M. & Streett, B. 1997: Impact-Echo. Jersey Shore, Bullbrier Press.

Scott, M., Rezai, A., Moore, M. & Washer, G. 2000. The HERMES Bridge Inspector Project, Reference Manual, Geophysics.

Streicher, D., Wiggenhauser, H., Holst, R. & Haardt, P. 2005. Zerstörungsfreie Prüfung im Bauwesen: Automatisierte Messungen mit Radar, Ultraschallecho und Impact-Echo an der Fuldatalbrücke. Beton- und Stahlbetonbau 100, 3: 216–224.

US Patent 2006/0191358 A1. www.acsys.ru 2006.www.olsoninstruments.com 2006.ZFPBau-Kompendium der Bundesanstalt für Material-

forschung und -prüfung; http://www.bam.de/service/publikationen/zfp_kompendium/welcome.html.